An efficient method for high-density multimodal salt-and-pepper noise removal of MRI images

Abstract

Medical image noise reduction is a significant and challenging area in image processing. A new adaptive window-based solution for the removal of high-density multimodal salt-and-pepper noise of the brain MRI images is proposed in this paper. In this efficient method, for each pixel of the noisy input image, an adaptive n x n window is considered in the neighborhood of that pixel, where n depends on the noise value. The higher the noise density, the larger the window size in which healthy pixels are found. If they are not noisy, the pixels of this window are weighted according to their distance from the desired pixel. The greater the distance, the less weight they gain. Then, the weighted sum of the neighboring pixels is averaged, and the noisy pixel replaces with the resulting value. To evaluate the proposed method against multimodal salt-and-pepper noise, which simultaneously appears in an image from 1–98%, 208 images from seven MRI databases are applied. The results show the excellent performance of the proposed method. The mean Peak Signal to Noise Ratio (PSNR) of whole databases is 29.3465. As a preprocessing step, the efficient proposed method shows highly accurate results on the brain MRI images. After applying the noise removal method, the quality and the Structural Similarity (SSIM) increased. In this study, in addition to removing multimodal noise in an image, noise with a specific density (single mode) in each image is also removed with similar or better results.